<8>

The most frequent strategy for focalizing a term (subject/object NP (or parts of it) or adverbial) in Fulfulde involves a bi-clausal structure being its first feature:

The first part contains the focalized element and will therefore be called in the following “focus clause”, the second one provides as “out-of-focus clause” the background information. Two aspects define the focus clause:

The fourth characteristic of the term focus construction is located in the out-of-focus clause:

<9>

The two verb forms are identical to those used for relative and other background clauses and will be discussed in detail below. The word order in the out-of-focus clause stays canonical: being an SVO language, prepositional phrases and manner adverbs are preferred sentence-finally, whereas time adverbs may be also placed sentence-initially (Diallo 2000). It is important to point out that the element which is focalized is not repeated in the out-of-focus clause.

On the suprasegmental level, there is a change of what Diallo & Ermisch (forth.) call the sentence melody (see also Anyanwu & Diallo 2005), being the fifth feature:

<10>

The prosody of focus constructions is thus different from neutral affirmative sentences where stress is placed on the verb root. In the following, I will discuss the morphosyntactic marking in more detail; hence no attention will be paid on the prosodic marking (see e.g. Anyanwu & Diallo 2005 or Bao 2012, the latter working on subject focus in the Senegalese variety).

The characteristics of the ex-situ term focus construction can be schematized as follows:

<11>

In the focus clause, we have the term focus marker ko and the focalized term X. Note that focalized pronouns will stand in their emphatic form. This first part of the bipartite structure is followed by the out-of-focus clause consisting of the subject S, the verb V, the object O and other elements Z (adverbs, prepositional phrases, etc.); the brackets indicate that once the subject, object or adverbial is focalized, it leaves no morphological trace in the out-of-focus clause.

<12>

Before going into detail regarding the verb form and the elements which can occur in the focus clause preceded by the term focus marker, I want to exemplify the described observations so far. Example (3) illustrates a neutral declarative sentence:

The sentence displays the canonical word order SVO, the active verb is conjugated in the imperfective 3 (the numbering will be explained shortly).

If the object is now in focus, e.g. after the question “What will Maria buy?”, the ex-situ structure will be:

The object occurs now in sentence-initial position preceded by the term focus marker, the second part of the clause maintains the canonical word order leaving a gap in the canonical object position. The verb is still conjugated in active voice, but now in the verb form imperfective 4.

<13>

As the further analysis of the construction will be partly based on the verb forms, it is necessary to look at them more closely. The verb system of Fulfulde is rather based on aspect than on tense; therefore in the literature three perfective verb forms are opposed to four imperfective verb forms. This asymmetry is due to the fact that in the traditional Fulfulde literature moods such as imperative and subjunctive are counted among the imperfective paradigm. In order to distinguish the verb forms within the perfective or imperfective, they are often simply numbered, although the number does not imply a specific meaning. Besides the division between perfective and imperfective, verbs may be conjugated in three different voices: active (a), middle (midd) and passive (pass).

The only verb forms permitted in the ex-situ construction are the forms perfective 2 and imperfective 4, having the following morphological shape in active, middle and passive voice:

<14>

In contrast to all other verb forms, the short subject pronouns of the 1^st person singular (only in a few regions of the Fuuta Jaloo accepted), 2^nd person singular, 1^st person plural inclusive and 2^nd person plural obligatorily follow the verb complex (i.e. verb root+TAM-suffix) in the perfective 2 and imperfective 4 modifying phonologically the verb forms above:

The encliticized pronouns induce a vowel change in the active forms (in the imperfective every enclitic pronoun triggers an own form, for details see Diallo 2000:171), in the middle and passive voice they lead to a vowel shortening (only the 1^st person singular enclitic does not evoke this phonological process in the imperfective). Note that the same changes happen when the preterite marker -no(o) is suffixed to the verb.

<15>

Example (5) illustrates this process for the short subject pronoun of the 3^rd person singular which never cliticizes and the subject pronoun of the 2^nd person singular which is obligatorily cliticized inducing the vowel change of the TAM-suffix:

The two verb forms perfective 2 and imperfective 4 are not only restricted to the usage in term focus constructions, but they occur also in narration, with stative verbs, in relative clauses, pseudo-clefts, temporal clauses and interrogatives.

<16>

Interrogatives are a good starting point to examine what kind of terms can be focalized in the focus clause as wh-questions in Fulfulde make use of the same ex-situ construction. The question word which is inherently in focus stands sentence-initial and is preceded by the term focus marker ko, the verb forms in the out-of-focus clause are restricted to the perfective 2 and imperfective 4 as described above. Hence, wh-questions can be schematized as in Figure 2:

Apart from wh-questions, objects and adverbials, the ex-situ construction is used for focus on the nominal and pronominal subject, illustrated in example (6):

<17>

In example (6A₁), the subject is a proper name, but can be replaced by any NP. The length of this NP is not restricted: The focalized subject NP may be indefinite or definite, a proper name, a pronoun or may be enlarged by a quantifier, qualifier or a (pro-)nominal possessor. When more than one word is contained in the NP, e.g. a noun and a quantifier, the whole NP or only part of it is in the scope of the term focus marker. This means that ex-situ focus constructions with larger NPs in the focus clause are potentially ambiguous, because only a part of the NP is semantically in focus while the whole NP is marked. Examples (7) and (8) illustrate this ambiguity for subject and object focus:

The ex-situ construction is also used for focalizing adverbial phrases:

<18>

To finish this overview of elements that can be focalized ex-situ, I want to give two last examples: Comparable to the examples (7) and (8) above, longer terms as prepositional phrases (example 11) and adverbial clauses (example 12) can be focalized ex-situ. Therefore, the whole phrase or clause is in sentence-initial position preceded by ko:

<19>

The data shown so far represent always sentences with one focalized term. But how are sentences with two semantic foci structured? The question/answer-pair in example (13) illustrates such a semantic multiple focus:

<20>

Although semantically there are two foci – one on the subject and one on the VP – only the term can be marked syntactically and morphologically. Field work data reveals that if one of these foci is on the subject, this one will be chosen for the sentence-initial position preserving the canonical word order, as in example (13). This observation can be explained by the fact that an unmarked subject is usually interpreted as the sentence topic, while an unmarked object is more likely interpreted as focal information. Hence, when the subject is not topical but focal, this needs to be marked additionally (see Fiedler et al. 2010).

We can now summarize the observations made on the ex-situ term focus construction in this section:

Apart from the observation that the ex-situ term focus construction is bi-clausal, no deeper analysis on its structure has been made so far. In the next section I will discuss the possibility of its analysis as a cleft construction.

<24>

In the previous section I have argued that the ex-situ term focus construction in Fulfulde of Fuuta Jaloo cannot be analyzed synchronically as a cleft sentence. The presented similarities to “real cleft” constructions (in Lambrecht’s terms) still raise the question of an historic link between the two constructions.

At this point the grammaticalization theory represented by Heine & Reh (1983, 1984) who worked on African languages offers an explanation for the resemblances of clefts and assertive term focus constructions. Discussing similarities and differences of relative and out-of-focus clauses in languages such as Akan (Kwa < Niger-Congo), Kikuyu or Duala (both Bantu < Niger-Congo), the authors conclude that cleft constructions are the source of morphosyntactic term focus marking in these languages. In some cases these systems are identical to clefts; in others they are just similar. In order to capture the differences between the structures observed, two types of morphosyntactic term focus marking systems are distinguished: weakly and strongly grammaticalized systems. The categorization depends on the degree of grammaticalization.

The structure of ex-situ term focus in Fulfulde discussed above would be counted amongst the weakly grammaticalized systems. These are characterized by two features:

The assumption is thus that the term focus construction is derived from a former cleft construction through a grammaticalization process. Adopting this hypothesis for the ex-situ term focus construction in Fulfulde of Fuuta Jaloo, this process would have taken place both in the focus clause and in the out-of-focus clause in three steps:

<25>

In order to apply the grammaticalization hypothesis to Fulfulde, let us look first at the focus clause. As described above, one can assume the grammaticalization of a copula sentence towards the focus clause. In Fulfulde this analysis is synchronically still transparent, as both lexemes – the copula (cop) and the term focus marker (tf) – are homophones, represented by the morpheme ko:

The function of both ko being different, they must be treated as distinct elements and thus be glossed differently. The grammaticalization path for the development of the term focus marker is illustrated in Figure 3 below:

<26>

The second feature of a weakly grammaticalized system was said to be the resemblance between the out-of-focus clause and a dependent, i.e. relative clause. As I discussed in section 4, the similarity between ex-situ focus constructions and relative clauses consists in the restriction to the same verb forms used in both cases; the difference lies in the absence of the relative pronoun in the former. As described above, a cleft sentence according to Lambrecht (2001) would contain a relative pronoun in the out-of-focus clause while the verb form would not change. In a second step, the cleft construction grammaticalizes towards the ex-situ term focus construction. During this process, a) the copula is reanalyzed as term focus marker, b) the relative pronoun is deleted and c) the definite article of the focalized term precedes the out-of-focus clause:

The cleft sentence in example (19) is not attested today in the Fulfulde of Fuuta Jaloo and is thus marked as ungrammatical. Since cleft sentences inherently promote the clefted element, the grammaticalized cleft construction is naturally open for other terms such as adverbials to occur in the focus clause, although those cannot function as the head of a relative clause.

<27>

Comparing examples (19) and (20), the question arises what drove the deletion of the relative pronoun in the out-of-focus clause, while it remained in “normal” relative clauses. In fact, the deletion within relative clauses is excluded in Fulfulde of Fuuta Diallo, but in its close variety of Maasina (Mali) the relative pronoun and the final definite article are actually optional in object relatives, as shown in example (21a):

Example (21b) illustrates that both the relative pronoun and the definite article are not deletable in Fulfulde of Fuuta Jaloo, although the deletion itself is not an unusual behavior of relative pronouns in other Fulfulde dialects.

In this subsection, I demonstrated the possibility of the grammaticalization of a former cleft construction (which is no longer attested in the Fulfulde of Fuuta Jaloo) to the present-day ex-situ term focus construction, following cross-linguistic observations of Heine & Reh (1983, 1984). During the grammaticalization process the deletion of the relative pronoun took place while the verb form remained. Also the class of constituents functioning as focus was extended in contrast to those constituents functioning as the head of a relative clause. The resemblances between focus and relative constructions may, according to Heine & Reh (1983), be explained by the function of the two types, i.e. the promotion of a constituent from the embedded to the matrix clause, leading to a partly parallel but diachronically independent development (Heine & Reh 1983, Bearth 1999). In the next section I will discuss the idea that a cleft sentence is not necessarily the underlying form of the term focus construction.

<28>

As already foreshadowed above, it is also possible that not a cleft sentence is the source of the ex-situ term focus construction, but rather that the language reanalyzed for the focus clause and the out-of-focus clause other linguistic devices that already existed in the language. Here again, I want to examine both the focus clause and the out-of-focus clause separately.

Instead of supposing that the term focus marker developed from a copula, another development could have taken place, as illustrated below:

The identificational marker is homophonous to the copula and the term focus marker:

Note that beyond the presented usages, ko is also used in other contexts, e.g. as a relative pronoun, making this morpheme highly polyfunctional in Fulfulde (for other usages see e.g. Diallo & Ermisch forth. or Evans 2001).

In example (22) we saw that the identificational marker ko identifies the following NP, directing the hearer’s attention towards the semantics of the latter representing thus the focal information:

<29>

In an intermediate step, a topic preceded the identificational sentence before the structure became a normal copula sentence ([ko NP]→[Topic, ko NP]→[NP ko NP]) where the identificational marker is reanalyzed as a copula. On information-structural level, the first NP of a copula sentence is interpreted as the topic of the utterance, the information status of the NP preceded by the copula being higher, i.e. in focus:

In examples (23) and (24), the morpheme ko promotes the following NP. The same happens in term focus constructions: in an intermediate step the identificational sentence is this time followed by a background clause ([ko NP] → [ko NP, background]). This became the ex-situ term focus construction [ko X background], where the identificational marker is reanalyzed as term focus marker. It promotes the term following it, preceding a background clause:

The upper examples show that instead of assuming a former cleft construction, it is also possible that the resemblance between identificational and focus clause arises from the same function they share, i.e. promoting the following element. In term focus constructions this element is not necessarily a NP.

<30>

Evans (2001) supports the idea in seeing the basic function of ko in the identification: the ex-situ subject focus construction is also used for thetic utterances (also called all-new sentences or sentence focus), e.g. in the first sentence of a narration:

The meaning of example (26) is thus ambiguous between subject focus and a thetic utterance.

Let us now look again at the out-of-focus clause in the focus construction. Instead of assuming that the verb forms which are otherwise used in relative clauses entered in the term focus construction via a cleft, one could also imagine that the verb system possessed a conjugation which was restricted to dependent background clauses and which became used in different background contexts, as illustrated below:

<31>

The function of this conjugation pattern – i.e. the backgrounding of the respective clause towards the main clause – would provide a reasonable explanation for the fact that always the same verb forms are used in relative clauses (as described above), in temporal clauses and also in term focus constructions. On these grounds it is not surprising that the shared features of relative verbal forms and those used in term focus constructions are a widespread phenomenon, being mentioned e.g. for English (Germanic < Indo-European), Hausa (Chadic < Afro-Asiatic) or Ilongot (Malayo-Polynesian < Austronesian) (see Schachter 1973, Zima 2006).

The importance of this backgrounding function by the verb form is more striking in other dialects of Fulfulde which use a “reduced” structure to express term focus. There, only the backgrounding verb form together with the syntactic fronting are used for marking term focus. In both examples (27) from the Malian dialect of Maasina and (28) from the Nigerian dialect of Gombe, the adverbial stands sentence-initially without the term focus marker ko:

Especially the last two examples demonstrate the value of the backgrounding feature of the out-of-focus clause promoting inherently the focalized term of the matrix clause. Although the morphological marking of the background of an utterance instead of its focus part is a less frequently described strategy for term focus (see e.g. Bagirmi (Central Sudanic<Nilo-Saharan), Jacob 2010), its function helps to understand the resemblances of relative and out-of-focus clauses.